Optical Fiber Angle Sensors for the PrHand Prosthesis: Development and Application in Grasp Types Recognition with Machine Learning

Abstract

This work presents the instrumentation of the PrHand upper-limb prosthesis with optical fiber sensors to measure the angle of the proximal interphalangeal joint. The angle sensors are based on bending-induced loss and are fabricated with polymer optical fiber (POF). The finger angle information is used in a k-Nearest Neighbor (k-NN) machine learning algorithm for grasp recognition. Four kinds of grasp are evaluated: hook grip, spherical grip, tripod pinch, and cylindrical grip, with three objects each. As mentioned in the algorithm validation, it is essential to note: The average accuracy was 92.81 %.